KR102395164B1 - Method and apparatus for providing speech based conversation service - Google Patents

Abstract

A method for providing a voice-based conversation service is provided. A method of providing a voice-based conversation service according to an embodiment of the present invention includes: acquiring an audio input from a user device; identifying the meaning of the audio input; and based on a determination that the identification of the meaning has failed , receiving the meaning from an administrator of the user device, and storing the received meaning.

Description

Method and apparatus for providing voice-based conversation service

The present invention relates to a method and apparatus for providing a voice-based conversation service. More particularly, it relates to a method and apparatus for efficiently providing a voice-based conversation service for infants or preschoolers whose language proficiency level is lower than that of adults.

The development of machine learning technology based on artificial neural networks has a great impact on the development of technologies that provide appropriate answers to questions and various related information through conversations with people, such as voice-based conversation services and text-based chatbot services. . In particular, machine learning technology promotes the development of detailed component technologies such as text generation, natural language processing, speech recognition, and speech synthesis, and devices such as artificial intelligence speakers that provide voice assistant services or voice-based conversation services They are contributing greatly to the spread of these in real life.

In particular, voice assistants and voice-based conversation services go beyond a function of handling a user's request, such as providing information upon a user's request or controlling another device according to a user's request to operate in a user's desired manner. It is developing to a level that provides functions that satisfy the needs of users who have not expressed them. For example, without an explicit request from the user, information that the user needs to know is provided in a timely manner, or entertainment content or emotional service to alleviate the boredom or loneliness of the user.

On the other hand, although the performance of the speech recognition natural language processing model based on machine learning may vary depending on the learning level of a specific language and a specific person's speech, the recognition rate for adult speech has been greatly improved based on recent major languages in the world. It has reached the level of providing a fairly natural service for a subject field that is used frequently.

On the other hand, there are still several difficulties in providing voice-based services to children, including infants and toddlers, who are not adults. Children have incomplete pronunciation because their vocal organs are not fully developed, their vocabulary and language skills are inferior compared to adults, and sometimes they use peer vocabulary that adults do not use or their own vocabulary that other children do not use. have This is a cause of not being able to provide a smooth voice-based conversation service by correctly recognizing a child's utterance and accurately grasping the meaning inherent therein.

In particular, since young children's utterances and vocabulary used by young children are often difficult to understand by anyone other than their parents or guardians or family members living together, the voice-based conversation service provider's side of the conversation log collected through the field service Even if humans listen and recognize the child's utterances included in the song and try to supervise the conversation engine, it is often difficult to understand the intention expressed by the child who is the speaker.

However, no technology or service to solve the above problems has been provided so far.

Korean Patent Publication No. 10-2017-0034154 (published on March 28, 2017)

A technical problem to be solved through some embodiments of the present invention is to provide a method and apparatus for efficiently providing a voice-based conversation service for children.

Another technical problem to be solved through some embodiments of the present invention is a method and apparatus for providing a voice-based conversation service capable of providing a response by properly understanding the meaning of the child's utterance and the vocabulary used by the child. will provide

Another technical problem to be solved through some embodiments of the present invention is to provide a method and apparatus for providing a voice-based conversation service capable of increasing the service use time of children by forming high intimacy with children who are users. .

Another technical problem to be solved through some embodiments of the present invention is to provide a method and apparatus for providing a voice-based conversation service capable of improving the vocabulary use habit of a child who is a user and notifying a guardian.

The technical problems of the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the following description.

In order to solve the above technical problem, a method for providing a voice-based conversation service according to an embodiment of the present invention includes: obtaining an audio input from a user device; identifying the meaning of the audio input; and identifying the meaning and receiving the meaning from an administrator of the user device based on a determination that the operation has failed, and storing the received meaning.

In an embodiment, the step of identifying the meaning of the audio input includes comparing the user uttered word recognized from the audio input with words contained in a dictionary included in a dictionary, and storing the received meaning includes adding the user uttered word to the dictionary.

In an embodiment, the comparing the user-spoken word with the words included in the dictionary includes comparing the user-spoken word with words included in a first dictionary, and personalizing the user-spoken word to the user device and comparing the words with words included in a second dictionary, wherein the adding of the user-spoken word to the dictionary includes adding the user-spoken word to the second dictionary.

In an embodiment, in the method for providing a voice-based conversation service, the step of comparing the user's spoken word with the words included in the dictionary includes comparing the user's spoken word with a third dictionary personalized to another user device that is distinct from the user device It may further include the step of comparing with the words included in the.

In an embodiment, the speaker of the other user device is a user belonging to the same speaker group as the speaker of the user device, and the speaker group may be determined based on at least one of an age and a spoken language level of the speaker.

In an embodiment, the requesting of the manager for information related to the user spoken word includes providing at least a part of a conversation history generated through the user device to the manager; Including the step of receiving input.

In an embodiment, the requesting of the administrator for information related to the user's spoken word may include: selecting a word matching the user's spoken word from among words included in a personalized dictionary in another user device that is distinct from the user device and providing it to the administrator.

In an embodiment, the requesting of the administrator for information related to the user uttered word may include: based on a determination that the speaker of the user device has uttered the user uttered word more than a preset number of times, the user uttered word and requesting information related to the manager.

In an embodiment, the step of requesting the information related to the user's spoken word from the manager includes receiving an input of whether the voice-based conversation service induces the speaker of the user device to correct the user's spoken word into a dictionary-recorded word. includes steps.

In one embodiment, based on a determination that the identification of the meaning is successful, setting the voice-based conversation service to use the user uttered word in a response utterance provided to the user device.

In one embodiment, on the basis of a determination that the identification of the meaning is successful, the voice-based conversation service sets to provide a word corresponding to the meaning in the content provided to the user device by replacing the word uttered by the user includes steps.

In one embodiment, the method for providing a voice-based conversation service includes determining a level of a speaker of the user device based on a determination that the identification of the meaning is successful, and determining that the speaker corresponds to a first level setting to provide a voice-based conversation service to the user device using the user uttered words based on the and setting the user device to provide a response utterance prompting to correct the user uttered word with the word.

In an embodiment, the method for providing a voice-based conversation service includes providing a voice-based conversation service to the user device by mixing the user uttered word and the matched dictionary entry on the basis of determining that the speaker corresponds to a third level Further comprising the step of setting to provide.

In one embodiment, determining the level of the speaker includes determining the level based on at least one of an age and a spoken language level of the speaker.

In an embodiment, the method for providing a voice-based conversation service provides information about the frequency with which the speaker of the user device uses the user uttered word and a dictionary-listed word matching the user uttered word for a preset period, respectively, to the user. The method further comprises providing to an administrator of the device.

In one embodiment, the step of identifying the meaning of the audio input comprises recognizing a user uttered word from the audio input, and based on a determination that the identification of the meaning has failed, to an administrator of the user device The receiving of the meaning may include providing the audio input to a manager of the user device, and the storing of the received meaning may include updating an audio recognition model based on information received from the manager. includes steps.

In an apparatus for providing a voice-based conversation service according to another embodiment of the present invention for solving the above technical problem, the steps of obtaining an audio input from a user device, identifying the meaning of the audio input, and identifying the meaning Based on the determination that the operation has failed, the steps of receiving the meaning from the manager of the user device and storing the received meaning are performed.

A computer-readable non-transitory recording medium according to another embodiment of the present invention for solving the above technical problem includes the steps of obtaining an audio input from a user device, identifying the meaning of the audio input, and the meaning A computer program is stored that causes the computer to perform a method comprising the steps of: receiving the input from an administrator of the user device, and storing the received meaning, based on a determination that the identification of the user has failed.

1 is a diagram illustrating an exemplary system in which a voice-based conversation service can be provided according to an embodiment of the present invention.
2 is a block diagram of an apparatus for providing a voice-based conversation service according to another embodiment of the present invention.
3 is a flowchart of a method for providing a voice-based conversation service according to another embodiment of the present invention.
FIG. 4 is a diagram for explaining in more detail some steps of the method of providing a voice-based conversation service described with reference to FIG. 3 .
5 and 6 are diagrams for explaining in more detail some steps described with reference to FIG. 4 .
FIG. 7 is a diagram for describing in more detail some steps described with reference to FIG. 6 .
8 and 9 are diagrams for explaining in more detail some steps described with reference to FIG. 4 .
10 is a diagram for explaining a hardware configuration of an apparatus for providing a voice-based conversation service according to some embodiments of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. Advantages and features of the present invention and methods of achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the technical spirit of the present invention is not limited to the following embodiments, but may be implemented in various different forms, and only the following embodiments complete the technical spirit of the present invention, and in the technical field to which the present invention belongs It is provided to fully inform those of ordinary skill in the art of the scope of the present invention, and the technical spirit of the present invention is only defined by the scope of the claims.

In adding reference numerals to the components of each drawing, it should be noted that the same components are given the same reference numerals as much as possible even though they are indicated on different drawings. In addition, in describing the present invention, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present invention, the detailed description thereof will be omitted.

Unless otherwise defined, all terms (including technical and scientific terms) used herein may be used with the meaning commonly understood by those of ordinary skill in the art to which the present invention belongs. In addition, terms defined in a commonly used dictionary are not to be interpreted ideally or excessively unless clearly defined in particular. The terminology used herein is for the purpose of describing the embodiments and is not intended to limit the present invention. In this specification, the singular also includes the plural unless specifically stated otherwise in the phrase.

In addition, in describing the components of the present invention, terms such as first, second, A, B, (a), (b), etc. may be used. These terms are only for distinguishing the elements from other elements, and the essence, order, or order of the elements are not limited by the terms. When it is described that a component is “connected”, “coupled” or “connected” to another component, the component may be directly connected or connected to the other component, but another component is formed between each component. It should be understood that elements may also be “connected,” “coupled,” or “connected.”

As used herein, "comprises" and/or "comprising" refers to the presence of one or more other components, steps, operations and/or elements mentioned. or addition is not excluded.

Hereinafter, some embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating an exemplary system in which a voice-based conversation service can be provided according to an embodiment of the present invention.

As shown in FIG. 1 , a system for providing a voice-based conversation service includes a voice-based conversation service providing server 1 (hereinafter referred to as a “server”) and a plurality of user devices 11a to 11d. In an embodiment, the voice-based conversation service providing system may further include a separate manager device 15 for managing user devices and/or an external content providing device 3 connected to the server 1 .

In this embodiment, the user devices 11a to 11d are, for example, smart devices, toys or dolls, smart watches, smart displays, and other home appliances that are equipped with artificial intelligence speakers, voice assistants or virtual assistant functions and have audio input/output functions, etc. includes The user devices 11a to 11d may perform voice-based interaction with the user through an audio input/output device such as a speaker and a microphone, and may additionally use an image input/output device such as a camera and a display device. The user devices 11a to 11d may transmit audio or video data input from the user to the server 1, and obtain responses or contents in various formats including audio or video from the server 1 and provide them to the user. there is.

In the present embodiment, the server 1 analyzes audio or video data received from the user devices 11a to 11d, and provides appropriate responses to the user devices 11a to 11d. In this case, the server 1 may receive content from a separate content providing device 3 managed by an external content provider and provide it to the user devices 11a to 11d in addition to the content it owns. The server 1 may include various detailed configurations to analyze the input from the user devices 11a to 11d and provide an appropriate conversation service based on the result, which will be described in more detail with reference to FIG. 2 . do it with

In the present embodiment, the server 1 and the user devices 11a to 11d are a local area network (LAN), a wireless network (Wireless Network), a wide area network (WAN), and a mobile network (Mobile Network). ) can be interconnected through

In one embodiment, the user devices 11a to 11d are short-range wireless communication such as Wi-Fi, Bluetooth, Bluetooth Low Energy (BLE), Zigbee, Z-Wave, RFID, NFC, etc. After being connected to a device capable of mid-to-long-range wireless communication such as a smart phone and a tablet PC (eg, the manager device 15), it may be connected to the voice-based conversation service providing server 1 through the device. As described above, when only a communication module for short-range wireless communication is embedded in the user devices 11a to 11d, there is an advantage in that the manufacturing cost of the user devices 11a to 11d can be reduced.

Hereinafter, the configuration and operation of the apparatus 1 for providing a voice-based conversation service according to an embodiment of the present invention will be described with reference to FIG. 2 . The voice-based conversation service providing apparatus 1 may be, for example, the server 1 of the voice-based conversation service providing system described with reference to FIG. 1 .

As shown in FIG. 2 , the apparatus 1 for providing a voice-based conversation service according to the present embodiment includes a conversation engine 10 , a dictionary management unit 20 , a manager device interface 30 , a content management unit 40 , and report generation. It may include the unit 50 , the content DB 60 , the dictionary DB 70 , and the conversation history DB 80 .

However, only the components related to the embodiment of the present invention are illustrated in FIG. 2 . Accordingly, one of ordinary skill in the art to which the present invention pertains can see that other general-purpose components other than those shown in FIG. 2 may be further included. In addition, each component of the apparatus 1 for providing a voice-based conversation service shown in FIG. 2 represents functional elements that are functionally separated, and a plurality of components may be implemented in a form that is integrated with each other in an actual physical environment. take note of Hereinafter, each component will be described.

The conversation engine 10 provides a function to process the overall conversation with the user device 11 . The conversation engine 10 is based on the data retrieved from the dictionary management unit 20 and the content management unit 40 and context information identified from the conversation history with the user device 11 recorded in the conversation history DB 80 , the user device (11) handle the conversation with. More specifically, it recognizes and analyzes user speech audio input from the user device 11 , and generates and provides output data such as an appropriate response or content appropriate to the situation.

The dictionary management unit 20 manages the dictionary by inquiring a word from a dictionary used in a conversation service provided by the apparatus 1 for providing a voice-based conversation service, and adding a new word to the dictionary. The dictionary data may be recorded and stored in the dictionary DB 70 .

The dictionary management unit 20 grasps the meaning of the word recognized by the dialogue engine 10 from the user's speech audio using the dictionary DB 70 , and provides the result to the dialogue engine 10 . Also, when a word whose meaning cannot be grasped using the dictionary DB 70 occurs, the dictionary management unit 20 transmits the word to the manager device interface 30 to be described later, and passes the word through the manager of the user device 11 . Try to figure out the meaning of the word.

The dictionary manager 20 may use and manage a plurality of dictionaries. For example, the dictionary management unit 20 includes a general vocabulary dictionary, a common dictionary including only words matching vocabulary used in the conversation service or topics provided by the conversation service, and dictionaries personalized to each of a plurality of users or user devices. ) can be managed and used by

In this embodiment, the dictionary personalized to the user or the user device may be a record of a unique word used by a specific user and the meaning of the word.

The unique word used by a specific user may be, for example, a word commonly used by peers of a specific age in a specific age (character names appearing in content for children, latest abbreviations or slang, etc.).

A unique word used by a specific user may be a unique word used only by a specific user, but not commonly used by other users. For example, there may be users who refer to or pronounce ice cream as “A goong” or “Akim”, and there may be users who refer to their grandmother as “Nana” and their grandfather as “Baba”. As such, the vocabulary used by infants and young children in particular varies greatly from child to child, and the same word is often used with different meanings. Therefore, it is necessary to record and manage in advance personalized to a specific user or a user device used by a specific user.

In addition, proper nouns such as names of family members or companion animals may be recorded and managed in the dictionary personalized to the user or user device.

In this embodiment, the dictionary management unit 20 manages and uses the personalized dictionary for a specific user or a specific user device as described above, thereby inducing the user to have a higher affinity for the conversation service, and ultimately, the conversation service can improve user engagement with

Next, the manager device interface 30 provides information to the manager device 15 used by the manager of the specific user device 11 , and obtains information necessary for a conversation service from the manager device 15 . The manager device 15 is, for example, a device used by a guardian such as a parent of a child who is a user of the user device 11 . In this embodiment, the manager does not refer to a management person on the side of the voice-based conversation service provider, but a person who owns and/or manages the user device 11 used by the end user, and includes a parent or guardian of the end user. Note that it means

Specifically, the manager device interface 30 provides the conversation history of the user (speaker) of the user device 11 to the manager device 15 , and provides the meaning of the words used by the speaker of the user device 11 to the manager device 15 . ) can be entered from In addition, the manager device interface 30 provides an audio clip uttered by the speaker of the user device 11 to the manager device 15, and the intention of the speaker recognized through the audio clip or the meaning of the sound contained in the audio clip. An input may be received from the manager device 15 .

Further, the manager device interface 30 may provide the manager device 15 with a report indicating the language ability or linguistic habits of the user (speaker) of the user device 11 . For example, statistical data on the type and frequency of words used by the user (speaker) of the user device 11 may be provided to the manager device 15 .

The report generating unit 50 may generate a report indicating the language ability or language habit of the user (speaker) of each user device 11 from the conversation history DB 80 , and provide it to the manager device interface 30 . there is.

Next, the content management unit 40 provides content in response to a content request from the conversation engine 10 . In the present embodiment, the contents include audio contents such as audio books and podcasts, video contents, interactive multimedia contents, and the like. When the user of the user device 11 is a child, content specific to the child may be provided.

The content may be obtained from the content DB 60 of the voice-based conversation service device 1 . In an embodiment, the content may be obtained from an external content providing device 3 . In one embodiment, only index data including the relation between the main word and the content identifier is stored in the content DB 60 , and the content management unit 40 transmits the content provided from the content providing device 3 to the conversation engine 10 . can transmit

The content management unit 40 may appropriately process the content provided from the content DB 60 and/or the content providing device 3 before providing it to the dialog engine 10 . For example, in order to induce the user (speaker) of the user device 11 to have higher intimacy, the name of the main character appearing in the content may be substituted with the name of the speaker, or a general word included in the content (eg, " Ice cream") may be provided to the conversation engine 10 through processing of replacing the word (eg, "a goong") unique to the speaker included in the dictionary personalized in the user device 11 . In another embodiment, the processing may be processed by the dialog engine 10 .

Next, an operation of the apparatus 1 for providing a voice-based conversation service according to the present embodiment will be described.

First, a conversation with a speaker of the user device 11 is initiated by the conversation engine 10 , and audio containing the user's utterance is obtained from the user device 11 . The conversation engine 10 records at least a part of the conversation exchanged with the user device 11 in the conversation history DB 80 . The dialogue engine 10 analyzes the user's speech audio to identify words included in the user's speech audio. The conversation engine 10 provides the identified user uttered word to the dictionary management unit 20 .

The dictionary management unit 20 compares the words stored in the dictionary DB 70 with the user's spoken words to determine the meaning of the user's spoken words. In this case, the dictionary manager 20 may use and inquire one or more types of dictionaries as described above. For example, the dictionary manager 20 inquires a general vocabulary dictionary, a common dictionary, a dictionary personalized to the user device 11 , a dictionary personalized to a user device other than the user device 11 , etc. can understand the meaning of

If, in at least one of the dictionaries, there is a dictionary-listed word matching the user's spoken word, the dictionary management unit 20 may determine the meaning of the user's spoken word. In this case, the dictionary manager 20 provides the meaning of the user's spoken word to the dialog engine 10 . The conversation engine 10 continues to provide a voice-based conversation service based on the above meaning. For example, the conversation engine 10 receives the content matching the subject (eg, snacks) including the user's uttered word (“a goong” meaning ice cream) through the content management unit 40, and the user device 11 can be provided to In addition, the conversation engine 10 may continue the conversation with the user device 11 while using the user's spoken word (“a goong”) in a sentence. As another example, the dialog engine 10 provides the user device 11 with content processed in such a way that the portion where the general noun “ice cream” is used is replaced with “a goong” in the ready-made content, so that the user of the user device 11 It can make them feel that they are being provided with a more intimate and personalized conversation service.

When there is no dictionary-listed word matching the user's spoken word in at least one of the dictionaries, the dictionary management unit 20 may have failed primarily to determine the meaning of the user's spoken word.

In this case, the dictionary manager 20 causes the dialog engine 10 to provide the user device 11 with an utterance asking the meaning of the user uttered word, and induces the speaker of the user device 11 to answer, so that the user You can figure out the meaning of spoken words.

In addition, when a predetermined condition is satisfied (eg, when the cumulative number of use of a private user uttered word exceeds a threshold), the dictionary management unit 20 is configured to use the user utterance through the manager device 15 through the manager device interface 30 . You can figure out the meaning of words. Here, the manager is a person who owns and/or manages the user device 11 from which the user uttered words are obtained, and is a guardian of the speaker of the user device 11 , and the manager device 15 controls the device used by the manager. Note that this refers to The manager device interface 30 may provide, for example, a part of a conversation record including user speech words to the manager device 15 , and receive the meaning of the user speech words from the manager device 15 . The dictionary manager 20 may add the user uttered word and its meaning to the dictionary DB 70 based on an input from the manager device 15 .

So far, the configuration and operation of the apparatus 1 for providing a voice-based conversation service according to the present embodiment have been described with reference to FIG. 2 . Hereinafter, a method of providing a voice-based conversation service according to another embodiment of the present invention will be described with reference to FIGS. 3 to 9 .

3 is a flowchart of a method for providing a voice-based conversation service according to the present embodiment. However, this is only a preferred embodiment for achieving the object of the present invention, and it goes without saying that some steps may be added or deleted as needed.

Each step of the method for providing a voice-based conversation service shown in FIG. 3 may be performed, for example, by the apparatus 1 for providing a voice-based conversation service described with reference to FIG. 2 . In other words, each step of the method for providing a voice-based conversation service may be implemented with one or more instructions executed by a processor of a computing device such as the apparatus 1 for providing a voice-based conversation service. First steps included in the method for providing a voice-based conversation service may be performed by a first computing device, and second steps of the method may be performed by a second computing device. Hereinafter, it is assumed that each step of the method for providing a voice-based conversation service is performed by the apparatus 1 for providing a voice-based conversation service. However, the subject performing each step is merely an example, and the present invention is not limited by the following description, and for the convenience of explanation, the subject of the operation of some steps included in the method for providing a voice-based conversation service is omitted. it might be

Of course, the technical ideas of the embodiments described with reference to FIGS. 1 and 2 may be reflected in each operation of the method for providing a voice-based conversation service according to the present embodiment, even if not separately mentioned. Conversely, the technical idea reflected in each operation of the method for providing a voice-based conversation service according to the present embodiment may also be reflected in the configuration and operation of the apparatus 1 for providing a voice-based conversation service described with reference to FIG. 2 .

As shown in FIG. 3 , the method for providing a voice-based conversation service according to the present embodiment includes the steps of providing a greeting (S10), obtaining user speech audio from the user device (S20), and analyzing the user speech audio (S10). S30), and providing a response (S40).

First, in step S10 , the voice-based conversation service providing device 1 may provide a greeting to the user device 11 . For example, when the user device 11 is connected to the voice-based conversation service providing device 1 for the first time after shipment and the conversation service is started, a greeting appropriate for such a situation may be provided, and the user device 11 is today When it is initially driven or when it is restarted after being paused, an appropriate greeting reflecting various situations is provided, thereby increasing intimacy with the speaker of the user device 11 .

The conversation model can then be used to attempt to have a conversation about a topic that the speaker may be interested in. For example, "What's your favorite animal?", "What flavor of ice cream do you like the most?" It is possible to provide a question sentence that can induce a conversation by attracting the speaker's attention, such as.

In step S20, the user's speech audio is obtained from the user device. For example, an answer to the question sentence may be obtained from a child using the user device 11 .

In step S30, the user uttered audio obtained in step S20 is analyzed. In step S30, an operation such as converting the user's utterance audio into text, natural language processing to determine the sentence structure, and inquiring the meaning of the word included in the sentence in the dictionary, may be performed. Step S30 will be described in more detail with reference to FIGS. 4 to 9 .

In step S40 , an appropriate response may be provided to the user based on the result analyzed in step S30 . For example, the conversation may be continued using the conversation model, or content matching the user's interest identified based on the user's utterance may be provided to the user through the user device 11 .

Hereinafter, with reference to FIG. 4 , a process of analyzing the user's speech audio in step S30 of FIG. 3 will be described in more detail.

Referring to FIG. 4 , first, in step S31 , audio of a user's speech is converted into text (STT: Speech to Text), and in step S32, natural language processing (NLP) is processed on the converted text. Thus, a sentence included in the user speech audio is recognized as a collection of structured user speech words. Various techniques and methods known in the art may be used for STT and NLP processing, and the present invention is not limited to using any one technique or method. Steps S31 and S32 may be independently performed, for example, by the dialogue engine 10 of the voice-based conversation service providing apparatus 1, but an external service connected to the voice-based conversation service providing apparatus 1 by a network Alternatively, it may be performed using an external engine. Although not shown in FIG. 4 , in some embodiments of the present invention, when the process of step S31 or S32 fails, the matching failure processing routine S35 of step S35 may proceed.

In step S33, the user's spoken word recognized in the previous step is compared with words recorded in the dictionary to determine the meaning of the user's spoken word. In other words, it is determined whether there is a dictionary-listed word matching the user uttered word. Step S33 will be described later in more detail with reference to FIG. 5 .

In step S34, if there is no dictionary-listed word matching the user's spoken word, it branches to step S35, and if there is a dictionary-listed word matching the user's spoken word, it branches to step S36.

In steps S35 and S36, a routine for a case in which the user spoken word does not exist in the dictionary and a routine for a case in which the user spoken word exists in the dictionary are respectively performed. This will be described in more detail with reference to FIGS. 6 to 9 .

Hereinafter, with reference to FIG. 5 , a process of comparing the user's spoken word with the dictionary in step S33 of FIG. 4 will be described in more detail.

In the method of providing a voice-based conversation service according to the present embodiment, as in the embodiment described with reference to FIG. 2 , a plurality of dictionaries may be used. The plurality of dictionaries include, for example, a general vocabulary dictionary, a common dictionary including only words matching vocabulary used in the conversation service or topics provided in the conversation service, and dictionaries personalized to each of the plurality of users or user devices. include

In step S331, the user's spoken word is searched for in the general vocabulary dictionary and/or the common dictionary. In other words, the user uttered word is compared with words included in the general vocabulary dictionary and/or the common dictionary, and it is determined whether a matching dictionary entry word exists.

In step S332, a user uttered word is searched for in a dictionary personalized to the current user device, and it is determined whether a matching dictionary entry word exists. The dictionary personalized to the user device may be a dictionary personalized to the user device #1 or the speaker #1 of the conversation in progress. The dictionary personalized to the user device may be a dictionary in which words uniquely used by the user device #1 into which the user utterance word is input or the speaker #1 using the device is recorded.

In step S333, a user uttered word is searched for in a dictionary personalized to a user device other than the user device #1 (user device #2, #3, ...), and it is determined whether there is a matching dictionary entry word. That is, the user's uttered word is not transmitted to the user device #1 of the speaker #1, but to the user devices (devices #2, #3, ...) of other speakers (speakers #2, #3, ...). In the personalized dictionary, it is queried whether the user uttered word exists.

In this case, the personalized dictionary of another user device used may be a dictionary personalized to the devices of speakers belonging to the same speaker group as speaker #1 of the user device #1.

The speaker group may be a group determined based on the speaker's residential area, age, language proficiency level, content consumption taste, and the like. For example, if speaker #1 is a 34-month-old child residing in the metropolitan area and is a child corresponding to the first level among the language proficiency levels classified into the first to third levels, a similar age and similar language proficiency level For branch, a dictionary personalized to the device of another speaker living in the metropolitan area can be used preferentially. As another example, if speaker #1 is a 45-month-old child living in the metropolitan area and a child who enjoys watching Robocar Poli animations, it is personalized to the device used by a child who enjoys watching similar content with a similar residence and age. A dictionary may be used preferentially.

In other words, in order to understand the meaning of the word spoken by the speaker #1 of the user device #1, words spoken by another speaker having a residence, age, language proficiency level, and/or preference similar to the speaker #1 are referred to. , it is possible to increase the efficiency and accuracy of the process of recognizing the meaning of words spoken by the user.

In another embodiment, the personalized dictionary of another user device used in step S333 may be a dictionary personalized to the devices of speakers belonging to the same speaker group having similar content consumption tastes as the speaker #1 of the user device #1. there is.

In step S334, the results searched for in steps S331 to S333 are synthesized, and if there is at least one matched dictionary entry word, a finally matched dictionary entry word may be determined.

For example, if a dictionary entry matching a user utterance word is found in two or more dictionaries, a finally matched dictionary entry word may be determined according to priority. For example, a word from a dictionary personalized on a user device has the highest priority, a word from a common dictionary has the next priority, and a word from a dictionary personalized on another user device has the lowest priority can have That is, the word recorded in the user's own dictionary uttering the user's spoken word has the highest priority, and the meaning of the corresponding word is used.

Finally, the matched dictionary entry is provided to the matching success processing routine of step S36.

Meanwhile, in step S334 , it may be determined that there is no dictionary-listed word matching the user uttered word. In this case, the user uttered word is a word that does not exist at all in the dictionary managed by the apparatus 1 for providing a voice-based conversation service, and is a word whose meaning cannot be grasped. If it is determined in step S334 that there is no dictionary entry that matches the user uttered word, it branches from step S34 to step S35 and a matching failure processing routine proceeds.

Hereinafter, the matching failure processing routine performed in step S35 of FIG. 4 will be described in more detail with reference to FIGS. 6 and 7 . First refer to FIG. 6 .

If the meaning of the user's spoken word cannot be grasped, first, in step S351, the user device attempts to grasp the meaning. The description will be continued taking the case where the meaning of the word "akim" uttered by the speaker of the user device 11 cannot be grasped as an example.

In step S351 , in order to understand the meaning of “Achim”, a question may be asked from the speaker through the user device 11 . For example, by providing the question "Achim? Akim is a word I do not know, what does it mean?" to the speaker through the user device 11, an attempt may be made to directly obtain the meaning of "Achim" from the speaker.

In one embodiment, rather than asking the meaning of "Achim" as an open question, a candidate word indicating what "Achim" means is selected, and a question asking whether the word is correct is provided to the speaker through the user device 11 can do. For example, a dictionary personalized to a speaker #2 of a user device #2 other than the speaker of the user device 11 may include that “aschim” means “ice cream”. In this case, "ice cream", which means "askim", which has a shape and pronunciation similar to "akim", may be selected as a candidate word, and "akim? are you talking about ice cream?" By asking the speaker of the user device 11 in this way, it is possible to grasp the meaning of the user's spoken word in a more effective way.

If it succeeds in grasping the meaning of the user's spoken word through the user device 11 in step S351 (S352), it is recorded in the personalized dictionary in the user device 11, and proceeds to step S40 of FIG. It is possible to continue providing voice-based conversation services.

If it is not successful to grasp the meaning of the user's spoken word through the user device 11 in step S351 (S352), proceed to step S353 and try to understand the meaning of the user's spoken word through the manager of the user device . Step S353 will be described in more detail with reference to FIG. 7 .

Referring to FIG. 7 , the step S353 includes comparing the cumulative number of uses of the user uttered word with a threshold ( S3531 ), requesting information about the user uttered word from the manager of the user device ( S3532 ), and the user uttered word It may include detailed steps such as obtaining the meaning of and whether correction is desired from an administrator (S3533), and adding the user uttered word and its meaning to a personalized dictionary in the user device (S3534).

First, in step S3531 , it is determined whether the accumulated number of times the user uttered word is used in the user device 11 exceeds a preset threshold. In step S3532 to be described later, prior to requesting information on the user's spoken word from the manager of the user's device 11, it is checked whether the accumulated number of times the user's spoken word is used exceeds a threshold, thereby the speaker of the user's device 11 It is possible to prevent the inefficiency of asking the manager for information on words that are not repeatedly used or words that are mispronounced by mistake.

In step S3532, information about the user's spoken word is requested from the manager of the user device. As described above, the manager refers to a person who owns and/or manages the user device 11 from which the user's spoken word is obtained. In many cases, a guardian of a child who mainly uses the user device 11 may correspond to the manager of the user device 11 . The matching relationship between the user device 11 and the manager or between the speaker and the manager of the user device 11 may be recorded and used in a DB (not shown) in the voice-based conversation service providing device 1 .

In step S3532, the voice-based conversation service providing device 1 sends a message or a push notification to the manager device 15 used by the manager through the manager device interface 30, for example, so that the manager of the user device It is possible to request to input information about the user uttered word. The voice-based conversation service providing apparatus 1 provides at least a part of the conversation history made through the user device 11 through the UI of the manager device 15, and the user uttered words have It can induce the administrator to input the meaning.

In an embodiment, in step S3532 , a candidate word corresponding to the user's spoken word may be presented to the manager together with the conversation history. For example, if the meaning of the user's spoken word is recorded in a personalized dictionary in a user device other than the user device 11 (eg, a device used by another child), the meaning of the user device 11 is determined by the administrator of the user device 11 . By presenting to the user, a clue about the meaning of the user's spoken word may be provided to the manager. In this case, the other user device may be a device used by a speaker belonging to the same speaker group as the speaker of the user device 11 . That is, if there is any word used by a speaker having a residence, age, language proficiency level, content consumption taste, etc. similar to the speaker of the user device 11 , the user can refer to the meaning can present

In the step S3532, for example, "My child used the word 'a goong', do you know what the mother means? For reference, one of the other children of our age has the meaning of 'ice cream' and 'a goong' By providing the message "You're using ." to the manager through the manager device 15, the manager can input the meaning of the user's uttered word "Awkong".

In step S3533 , the meaning of the user uttered word may be acquired from the manager through the manager device 15 . By receiving the meaning of a word used by a speaker (eg, a child) of the user device 11 from an administrator (eg, a parent or guardian of the child) of the user device 11 , the parent of a child using the user device 11 or It is also possible to grasp the meaning of the unique expressions of a specific child that only the caregiver can know.

In an embodiment, in step S3533 , information on whether the administrator desires to induce the speaker of the user device 11 to correct the user uttered word as a standard word may additionally be obtained. Depending on the age or language proficiency level of the child who is the speaker of the user device 11 , it may be desirable to induce correction for continuing to use the speaker's own words, not the standard language, or continue to use the speaker's own words for a certain period of time. As to whether or not to use , the preferences of the parent of the child who is the speaker may be different. Also, the preference of the speaker's parent may be different depending on what the user's uttered word is and what the tone is. In step S3533 , an interface is provided for the administrator to specify whether or not to induce the user's spoken word to be corrected to a standard language, so that the voice-based conversation service affects the language usage habit of the speaker of the user device 11 . Administrators can decide whether to

In step S3534 , the meaning of the user uttered word received from the manager through the manager device 15 is recorded in the personalized dictionary in the user device 11 together with the user uttered word. When the same word is uttered from the user device 11 in the future, the meaning of the user uttered word can be grasped in the dictionary inquiry process of step S33 .

In some embodiments, in step S3532 , an audio clip including a user utterance other than the user utterance words is provided and played to the manager device 15 , and the meaning of the speaker's expression contained in the audio clip including the user utterance is displayed. You can have the administrator enter it. Since the speech of a child with incorrect pronunciation is often not recognized by the audio recognition model that converts audio into text, it may not be converted into user speech words in text form. In this case, by providing the audio clip itself including the user's utterance to the manager (the speaker's parent, etc.), even the meaning of the utterance that has not been converted into text can be grasped. In addition, based on the results, the audio recognition model can be supervised and updated.

So far, the matching failure processing routine performed in step S35 of FIG. 4 has been described with reference to FIGS. 6 and 7 . In the matching failure processing routine according to the present embodiment, the meaning of the user's uttered word for which a matching word is not found in the dictionary inquiry process is acquired from the manager or owner of the user device in which the utterance is made, such as the parent of the user who uttered the word. This has a particularly great effect when the target of the voice-based conversation service is a child, such as an infant. Since the utterances of young children and vocabulary used by young children are often difficult to understand by others other than parents, guardians, and family members living together, the meaning of the words spoken by the young child can be interpreted by the operator or the operator hired by the service provider. It is almost impossible for an administrator to find out and add it to the dictionary. In the embodiments of the present invention, since the meaning of the word is obtained from the administrator of the user device, that is, the parent, etc. in which the young child's utterance is made, the meaning of the word used by the young child is efficiently and accurately grasped, and a more advanced voice-based It has the advantage of being able to provide a conversation service.

Hereinafter, the matching success processing routine performed in step S36 of FIG. 4 will be described in more detail with reference to FIGS. 8 and 9 .

8 is a flowchart of a matching success processing routine according to some embodiments of the present invention. When a dictionary-listed word matching the user's spoken word is found from the dictionary, in step S36, a voice-based conversation service using the user's spoken word may be provided.

For example, referring to step S361 of FIG. 8 , a pre-recorded word matching the user's spoken word, that is, audio/video content corresponding to the meaning of the user's spoken word may be provided to the user device 11 . . For example, when it is determined that the user's uttered word is "a goong" and "a goong" matches "ice cream," audiobook content or video content related to "ice cream" may be provided through the user device 11 .

As another example, referring to step S362 of FIG. 8 , a user uttered word may be used in a response utterance of the dialogue engine based on the meaning of the word included in the dictionary. For example, "Which flavor do you like the most?", "When was the last time you ate a goong with your mom?" As such, by providing the user device 11 with a response utterance using “a goong” instead of “ice cream”, it is possible to form intimacy with the speaker of the user device 11 using the expression “a goong”.

As another example, it is possible to perform processing of replacing words in the dictionary appearing in the existing contents with words spoken by the user (step S3631), and provide the processed contents to the user device 11 (step S3632). For example, in a part where the word "ice cream" appears in the audiobook content of a children's story, "ice cream" may be replaced with "a goong", and the processed content may be provided to the user device 11 in this way. In this way, by allowing the voice-based conversation service providing device 1 to properly imitate the expression uniquely used by only the speaker of the user device 11, the user device 11’s intimacy with the speaker can be further improved, Ultimately, it will be possible to maximize user engagement with the conversation service.

In some embodiments, the report on the usage of the user uttered word may be provided to the administrator of the user device 11 (step S364). Specifically, the level of development of the speaker's language usage habit and language ability of the user device 11 by providing statistical data on the frequency of use, relative usage ratio, etc. of the user's spoken word and the corresponding standard word to the manager device 15 . , so that the speaker's parents, etc. can understand it.

9 is a flowchart of a matching success processing routine according to some other embodiments of the present invention. This embodiment is different from the embodiment described above with reference to FIG. 8 in that the response of the voice-based conversation service is different according to the language proficiency level of the speaker of the user device 11 .

Referring to FIG. 9 , first, the language proficiency level of the speaker of the user device 11 is determined (step S366).

When the speaker's language proficiency level is the first level, a voice-based conversation service is provided using the user's spoken words (step S367). For example, for a speaker with incomplete pronunciation and a low level of language proficiency due to not fully developed vocal organs, a voice-based conversation service is provided using the user's utterance word "a goong", which means "ice cream." It can promote intimacy. For a specific example of providing a voice-based conversation service using a user uttered word, reference may be made to descriptions of steps S361 to S3632 of the flowchart shown in FIG. 8 .

When the speaker's language proficiency level is the second level, a voice-based conversation service is provided using a dictionary-listed word (standard language, etc.) that matches the user's spoken word, and correction of the speaker's word can be induced (step S369) ). The second level may be a level that speaks a higher level language than the first level. For example, if the speaker's pronunciation is accurate enough that there is no problem in pronouncing "ice cream" and the speaker's language skills are sufficient, even if the speaker says "a goong", he responds using the standard word "ice cream" to encourage the speaker to use the standard word in the future. can induce

If the speaker's language proficiency level corresponds to a third level between the first level and the second level, a voice-based conversation service may be provided by mixing a user's spoken word and a dictionary-listed word (standard language, etc.). With respect to the speaker corresponding to the third level, it is possible to induce the speaker to gradually use the standard word by using the standard word while maintaining intimacy by using the user's spoken word.

So far, a method for providing a voice-based conversation service according to an embodiment of the present invention has been described with reference to FIGS. 3 to 9 .

According to this embodiment, by managing and using a dictionary personalized to a specific user or a specific user device, it is possible to induce a user to have a higher affinity for the conversation service and ultimately improve user engagement with the conversation service. be able to

Also, according to the present embodiment, by referring to a dictionary personalized to another user having a residence, age, language proficiency level, content consumption taste, etc. similar to the user, it is possible to effectively grasp the meaning of the word spoken by the user.

In addition, according to the present embodiment, with respect to a user's uttered word, the meaning of which cannot be grasped through a dictionary inquiry, by having the administrator of the user device or the user's parent input the meaning of the word, other People will be able to efficiently grasp the meaning of words spoken by young children, which are extremely difficult to understand. This makes it possible to grasp the meaning of words with a high level of accuracy that cannot be expected from the previous method of identifying the vocabulary of a specific child through a third party not related to the specific child, such as an operator or manager hired by the service provider. .

In addition, according to the present embodiment, in accordance with the user's language proficiency level or language development level, the user's uniquely used words and the corresponding standard words are appropriately used to provide a voice-based conversation service, thereby forming intimacy with the user. However, it is possible to gradually induce the improvement of the user's habit of using words.

In addition, according to the present embodiment, a report on the user's word usage habit is automatically generated and provided to the user's parents, so that the parent can easily understand the language proficiency level and developmental level of the child.

Hereinafter, with reference to FIG. 10 , an exemplary computing device 1500 capable of implementing the apparatus 1 for providing a voice-based conversation service in some embodiments of the present invention will be described.

10 is a hardware configuration diagram illustrating an exemplary computing device 1500 capable of implementing the apparatus 1 for providing a voice-based conversation service according to some embodiments of the present invention.

As shown in FIG. 10 , the computing device 1500 loads one or more processors 1510 , a bus 1550 , a communication interface 1570 , and a computer program 1591 executed by the processor 1510 . ) may include a memory 1530 and a storage 1590 for storing the computer program 1591 . However, only the components related to the embodiment of the present invention are illustrated in FIG. 10 . Accordingly, a person skilled in the art to which the present invention pertains can know that other general-purpose components other than the components shown in FIG. 10 may be further included.

The processor 1510 controls the overall operation of each component of the computing device 1500 . The processor 1510 includes a central processing unit (CPU), a micro processor unit (MPU), a micro controller unit (MCU), a graphic processing unit (GPU), or any type of processor well known in the art. can be In addition, the processor 1510 may perform an operation on at least one application or program for executing the method according to the embodiments of the present invention. The computing device 1500 may include one or more processors.

The memory 1530 stores various data, commands, and/or information. The memory 1530 may load one or more programs 1591 from the storage 1590 to execute a method according to embodiments of the present invention. The memory 1530 may be implemented as a volatile memory such as RAM, but the technical scope of the present invention is not limited thereto.

The bus 1550 provides a communication function between components of the computing device 1500 . The bus 1550 may be implemented as various types of buses, such as an address bus, a data bus, and a control bus.

The communication interface 1570 supports wired/wireless Internet communication of the computing device 1500 . Also, the communication interface 1570 may support various communication methods other than Internet communication. To this end, the communication interface 1570 may be configured to include a communication module well known in the art.

According to some embodiments, the communication interface 1570 may be omitted.

The storage 1590 may non-temporarily store the one or more programs 1591 and various data.

The storage 1590 is a non-volatile memory such as a read only memory (ROM), an erasable programmable ROM (EPROM), an electrically erasable programmable ROM (EEPROM), a flash memory, a hard disk, a removable disk, or well in the art to which the present invention pertains. It may be configured to include any known computer-readable recording medium.

The computer program 1591 may include one or more instructions that, when loaded into the memory 1530 , cause the processor 1510 to perform methods/operations in accordance with various embodiments of the present invention. That is, the processor 1510 may perform the methods/operations according to various embodiments of the present disclosure by executing the one or more instructions.

In this case, the devices according to some embodiments of the present invention may be implemented through the computing device 1500 .

So far, various embodiments of the present invention and effects according to the embodiments have been described with reference to FIGS. 1 to 10 . Effects according to the technical spirit of the present invention are not limited to the above-mentioned effects, and other effects not mentioned will be clearly understood by those skilled in the art from the following description.

The technical idea of the present invention described with reference to FIGS. 1 to 10 may be implemented as computer-readable codes on a computer-readable medium. The computer-readable recording medium may be, for example, a removable recording medium (CD, DVD, Blu-ray disk, USB storage device, removable hard disk) or a fixed recording medium (ROM, RAM, computer-equipped hard disk). can The computer program recorded on the computer-readable recording medium may be transmitted to another computing device through a network such as the Internet and installed in the other computing device, thereby being used in the other computing device.

In the above, even though it has been described that all components constituting the embodiment of the present invention are combined or operated in combination, the technical spirit of the present invention is not necessarily limited to this embodiment. That is, within the scope of the object of the present invention, all the components may operate by selectively combining one or more.

Although acts are shown in a particular order in the drawings, it should not be understood that the acts must be performed in the specific order or sequential order shown, or that all illustrated acts must be performed to obtain a desired result. In certain circumstances, multitasking and parallel processing may be advantageous. Moreover, the separation of the various components in the embodiments described above should not be construed as necessarily requiring such separation, and the described program components and systems may generally be integrated together into a single software product or packaged into multiple software products. It should be understood that there is

Although embodiments of the present invention have been described above with reference to the accompanying drawings, those of ordinary skill in the art to which the present invention pertains can practice the present invention in other specific forms without changing the technical spirit or essential features. can understand that there is Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. The protection scope of the present invention should be interpreted by the following claims, and all technical ideas within the equivalent range should be interpreted as being included in the scope of the technical ideas defined by the present invention.

Claims (18)

A method of providing a voice-based conversation service, comprising:
obtaining an audio input from a user device;
identifying the meaning of the audio input;
receiving the meaning from an administrator of the user device based on a determination that the identification of the meaning has failed; and
Storing the received meaning
including,
The step of identifying the meaning of the audio input comprises:
comparing the user uttered words recognized from the audio input with words included in a dictionary personalized to another user device distinct from the user device;
containing,
How to provide a voice-based conversation service. According to claim 1,
The step of identifying the meaning of the audio input comprises:
Comprising the step of comparing the user speech word recognized from the audio input with dictionary-listed words included in a dictionary personalized to the user device,
The step of storing the received meaning further includes adding the user uttered word to a personalized dictionary in the user device.
How to provide a voice-based conversation service. delete delete According to claim 1,
The speaker of the other user equipment is a user belonging to the same speaker group as the speaker of the user equipment,
The method for providing a voice-based conversation service, wherein the speaker group is determined based on at least one of an age and a spoken language level of the speaker. 3. The method of claim 2,
The step of receiving the meaning from the manager of the user device comprises:
providing at least a portion of a conversation history generated through the user device to the manager; and
Comprising the step of receiving the meaning of the user uttered word from the manager,
How to provide a voice-based conversation service. 3. The method of claim 2,
The step of receiving the meaning from the manager of the user device comprises:
and providing, to the administrator, a word matching the user uttered word among words included in a personalized dictionary to another user device, which is distinguished from the user device.
How to provide a voice-based conversation service. 3. The method of claim 2,
The step of receiving the meaning from the manager of the user device comprises:
based on a determination that the speaker of the user device has uttered the user uttered word more than a preset number of times, comprising the step of requesting information related to the user uttered word from the administrator,
How to provide a voice-based conversation service. 3. The method of claim 2,
The step of receiving the meaning from the manager of the user device comprises:
Receiving an input whether the voice-based conversation service induces the speaker of the user device to correct the user uttered word into a word recorded in the dictionary
containing,
How to provide a voice-based conversation service. 3. The method of claim 2,
Based on the judgment that the identification of the above meaning was successful,
and setting, by the voice-based conversation service, to use the user uttered word for a response utterance provided to the user device.
How to provide a voice-based conversation service. 3. The method of claim 2,
Based on the judgment that the identification of the above meaning was successful,
and setting a word corresponding to the meaning in the content provided to the user device by the voice-based conversation service to be provided by replacing the word uttered by the user,
How to provide a voice-based conversation service. A method of providing a voice-based conversation service, comprising:
obtaining an audio input from a user device;
identifying the meaning of the audio input;
receiving the meaning from an administrator of the user device based on a determination that the identification of the meaning has failed; and
Storing the received meaning
including,
determining a level of a speaker of the user device based on a determination that the semantic identification is successful;
setting, based on a determination that the speaker corresponds to a first level, to provide a voice-based conversation service to the user device using a user uttered word recognized from the audio input; and
setting, based on determining that the speaker corresponds to the second level, to provide the user device with a response utterance that prompts the user to correct the user uttered word with a dictionary entry matching the user uttered word;
further comprising,
How to provide a voice-based conversation service. 13. The method of claim 12,
setting, based on the determination that the speaker corresponds to the third level, to provide a voice-based conversation service to the user device by mixing the user uttered word and the matched dictionary entry word;
further comprising,
How to provide a voice-based conversation service. 13. The method of claim 12,
The step of determining the level of the speaker,
determining the level based on at least one of the age of the speaker and the level of language spoken;
How to provide a voice-based conversation service. 3. The method of claim 2,
providing information on the frequency with which the speaker of the user device uses the user uttered word and the dictionary-listed word matching the user uttered word to the manager of the user device, respectively, during a preset period
further comprising,
How to provide a voice-based conversation service. According to claim 1,
The step of identifying the meaning of the audio input comprises:
recognizing user-uttered words from the audio input;
Based on the determination that the identification of the meaning has failed, the step of receiving the meaning from an administrator of the user device includes:
providing the audio input to an administrator of the user device;
The step of storing the input meaning is,
Including the step of updating the audio recognition model based on the information received from the manager,
How to provide a voice-based conversation service. 17. An apparatus for providing a voice-based conversation service for performing the method according to any one of claims 1, 2 and 5 to 16. 17. A computer-readable non-transitory recording medium storing a computer program for causing a computer to perform the method for providing a voice-based conversation service according to any one of claims 1, 2, and 5 to 16.

Images